On the problem of the chemical synthesis of proteins with special reference to insulin

On the Problem of the Chemical Synthesis of Proteins with Special Reference to Insulin By P. G.

KATSOYAXXIS

HE introduction by San'ger in 1945 of the dinitrophenyl method for labeling free amino groups in a polypeptide chain, and the subsequent development of other tools for structural analysis led to the determination of the primary structure of an impressive number of polypeptides and proteins. Insulin, ACTH, oxytocin' and the vasopressins, MSH, glucagon, ribonuclease, and most recently, myoglobin, arc only a few examples among: the polypeptides and proteins whose structure has be en determined thus far. Furthermore, with the more recent advancements in degradative technics and the refinement of physical methods of analysis, it appears that elucidation of the structure of several more proteins is a matter of time only. 'Vith such impressive achievements, it appeared that protein structural analysis had outstripped the synthetic capabilities of peptide chemistry, the role of the latter supposedly being confined to the synthesis of small pcp tides, presumably to be used as enzyme substrates. Consequently, we came to believe, not too long ago, that the synthesis of polypeptides cotaining more than' about 10 amino acid residues was beyond our synthetic capacity and, of course, that the synthesis of even low molecular weight proteins was a dream of the future. However, an evaluation of the accomplishments of peptide chemistry in the past few years result's in a radically different picture. During this period, not only the synthesis of naturally occurring polypeptide sequences containing from nine to 24 amino acid residues became a reality, but also the synthesis of a considerable number of analogues of these polypeptides was accomplished. Because of these achievements, a better understanding of the relationship between chemical structure and the pharmacologican activity of several polypeptide hormones has been' attained and it app ears that this problem can now be pursued in an unlimited way. It will be beyond the scope 'of this discussion to consider in detail the events that made possible these great strides in peptide chemistry. I do believe, however, that a very brief account of the problems involved in the synthesis of polyp eptides arid the means we usc to cope with these problems will he a proper background for the synthetic studies that I propose to discuss with you. In undertaking the synthesis of a polypeptide chain, we face many problems. Among these, in our ex-perience, the following arc probably the most critical: 1: Need for protective groups that show different reactivities towards various deblocking reagents, for in the synthesis of polypeptides contain-

T

= =

Tile folloll:illg abbrc ciations arc IISCt! ill the figllrcs: Ae = accujl, B ill y-I Crl-lJUtyl b cn::.yl; Clrza carbobcnzoxo ; Z = carb oh cnzasij; IJ-NO 1.2 p-nitrocarboester; Bz bcnzosij; NB z p-nitrob cn::yl; To s = p-totucnesuljonu! (tosyl); DCC = N,Nl_clicyc1ohcxtjlcatbotlilmidc; PNP p-nitropllCnyl ester; CDI N,Nl-carbonyl dllmidazolc; Mc = methyl; R = alkyl; X carboxyl actir:atillg group,

=

=

=

=

=

=

1059 ~IETADol.I s:\r,

VOl.. 13, No. lO-PAnT 2 (OCTODEII), lO/H

1060

1'. G. KATSOYANNIS

in'g several protected functions, selective unmasking of a particular group is of prime importance for elongation of the peptide chain. 2. A second problem is the occurrence of racemization. With the exception of the azide method, in .whieh activation of the carboxyl group is accomplished by conversion to an azide, all the known condensation procedures are potentially capable of causing, at one time or another, partial or even complete racemization. 3. Finally, solubility problems and the question of yield of the product often hamper the synthesis of complex polypeptides. Solubility problems, particularly, which are involved not only in the actual synthesis but also in the purification of a peptide, are often one of the main road blocks for the successful construction of a polypeptidc chain. The situation regarding the availability of protective groups for, the reactive functions of the amino acids has improved enormously with recent developments. Thus, the introduction of the tert-butyloxycarbonyl, the trilluoroacetyl groups, and the revival of the trityl and formyl groups has greatly enriched the array of amino-protecting agents that can be cleaved selectively. Similarly, the use of tort-butyl esters and P: nitrobenzyl esters, along with the other Classical methods of protection, has increased considerably our capabilities for selective masking and unmasking of carboxylic functions. 1110 problem of racemization has not been solved as yet per se. Accumulated experience, however, indicates that even with those condensation procedures that arc highly prone to cause racemization, experimental conditions may be found that wiII keep the degree of racemization to a minimum, Furthermore, the availability of highly purified preparations of leucinaminopeptidase, carboxypeptidase, trypsin and chymotrypsin has provided an excellent tool for demonstrating the occurrence of racemization in the synthesis of a polypeptide. Complete digestion of the synthetic product by these enzymes to the constituent amino acids in the theoretically expected ratio is an adequate proof that only L-amino acid residues are present. Deamination of the amino acids of a hydrolysate of the synthetic peptide by L-amino acid oxidase, or lack of deamination with D-amino acid oxidase, leads to similar conclusions. Therefore, although we cannot always predict when and where racemization might occur, analytical tools are at least available for its defection and thus gUide our efforts in the correct synthetic approach. The question of yield and the solubility problem in peptide synthesis are often inseparable. If in a synthetic step there is a way to separate the product from the reactants without mechanical losses, obviously the overall yield depends on the peptide bond-forming reaction and therefore we could take full advantage of the efficiency of the peptide bond fonning method used in that particular step. Such an ideal case prevails in the preparation of small pep tides where solubility problems are practically non-existent. In the synthesis of large peptide fragments, however, the solubility of the reactants, large peptides themselves, and of the product is very similar and often very limited indeed. In such cases, because of the difficulties involved in the purification

PHOBLE~I

1061

OF ClIE},UCAL SYNTHESIS OF PHOTEINS

~I

N~ .CHCOOR R

IZ

~z

Z·NHCHCOOX .. R

I. de-acylation

II

Z· NH •CHCO-NH' CHCOOR 2. Z·NH' CHCOOX' I

R3 R3 1

~

I'"

R

II

I. de-acYlation

Z·NH· CHCO-NHCHCO-NHCHCOOR 2. Z'NH'CHCOOX' I

R4

R4 R3 RZ RI I 1 I I Z·NH· CHCO-NHCHCO-NHCHCO-NH'CHCOOR

Fig. I.-Stepwise elongation of the peptide chain.

of the product, the overall yield is considerably decreased no matter how efficient the peptide bond-forming method used. A partial solution of these problems is offered by application of the stepwise elongation of the peptide chain, adding onto the amino terminus of this chain. In practice the stepwise method consists in the condensation of an activated acylamino acid with a peptide ester. After each stepwise condensation the newly formed acylpeptide is deblocked to a new amino-free peptide ester which in turn is condensed with another acylamino acid, and the process is repeated until the desired polypeptide is synthesized. Figure 1 illustrates this approach. An amino acid or peptide ester is reacted with an activated acylamino acid to give an acylpeptide ester. This ester is de-acylated and the resulting new amino component is condensed with another acyl amino acid to give a new acylpeptide. In this approach the yields arc high since any of the most efficient peptide bond forming processes can be employed. The products arc of high optical purity since activation of an acylamino acid is usually not accompanied by racemization. Finally, the peptide constructed after each synthetic step can be readily purified, as might be expected by the fact that the solubility of the aeylpeptide ester formed is markedly different from the reactants, namely acylamino acid and amino-free peptide ester. In our synthetic studies we uscd almost exclusively this approach and we activated the acylamino acids used in each synthetic step by conversion to the corresponding ]J-nitrophenyl esters. As the amino protecting group in almost all the steps, we used the carbobenzoxy group. The major disadvantage of the stepwise approach is the great number of deblocking steps required for its execution. Deblocking of peptides, particularly as the length of the chain increases, creates time-consuming purification problems and augments the chance of side reactions. Thus, the stepwise method becomes more or less impractical when complex, peptide sequences

1062

1'. G. !i:ATSOYAKI'IS

I

5 5

15

18

2\

If> !G 7 8 9 10 II 12 \3 14 'fz 16 17 ~ 19 20 If: GIy·llou-VcI-G lu-Glu-Cy-Cy-AJa -Gly -Val-Cy -S0' -Lou - Tyf -Glu-Lou-Glu -Asp-Ty, -Cy -Asp I

A-Cho ;n

5

I

2

3

4

I

/

s 3

I

2

I

4

~ NI~

s

/

S 5

6

S

\7 8

9

10 II

12 13 14 15 16 17 18

h

20 2\

22 23 24

25

26 27 28 29 30

B - Choin PtIe -\\lI-Asp- Glu- His-Lou-Cy- Gly-So,-H;s- Lou-Val- Gru-Ala -Lo u-Ty,-Lou-Val-Cy-Gly- G'u-Arg -(;ly-P!l
Fig. 2.-The structure of sheep insulin .

containing more than 15 amino acid residues are to be constructed. In such cases the only alternative route for synthesis is .by "fragmcnt condensation." That is, preparation of peptide subunits by the stepwise approach, followed by their combination to the higher polypeptide. Obviously then, the proper selection of these subunits, and the choice of the correct methods for their condensation, is of key importance for the successful preparation of long polypeptides. It is apparent from this brief account of the "status quo" of peptide chemistry that we have not reached as yet the ideal situation necessary to cope with the synthesis of large protein' molecules. It is my belief, however, that the refinement of the synthetic methodology, and the de velopment of purification technics accomplished thus far, have set the stage for the synthesis of molecules approaching the size of low molecular weight proteins. With this belief we started the work that I am going to discuss with you today. 111is work was carried out during the last four years in our laboratory and is concerned with the synthesis of certain' structural clements of the insulin molecule and eventually with the total synthesis of this protein. Degradative studies hy Sangcr and co-workers led to the elucidation of the complete amino acid sequence and subsequently the overall structure of insulin from various species. In this structure, shown in' figure 2, two polypeptide chains are present: the A-chain containing 21 amino acid residues, with glycine as the N-terminus, and the B-chain containing 30 amino acid residues, with penylalaninc as the Nvtcrminal amino acid. In the insulin molecule, these two chains are linked together by two disulfide bridges located at positions 7 and 20 in the A-chain, and 7 and 19 in the Bvchain, In addition there is an intrachain disulfide bridge linking positions 6 and 11 in the A-chain which results in the formation of a 20·membered cystine-containing ring system. Minor variations in amino acid sequence occur among the insulins of different species. In undertaking the synthesis of insulin we were firmly convinced that the preparation of polypeptides of complexity and length similar to that of the insulin chains is attainable with the present synthetic methodology. Furthermore we made the assumption that if chemically synthesized A and B chains are available, it might he possible to obtain insulin by air-oxidation of a mixture of the sulfhydryl forms of the two chains. It appears now that we are fulIy justified in assuming that chemical synthesis of the A- and B-chains could lead to insulin' synthesis. Recently Dixon

1063

PHOllLE:\[ OF CHEMICAL SYNTHESIS OF PROTEINS

5-

-

-

-

I (7) NHZ'GIY ---Cys'Cys I (I)

(6)

-

5

(II)C

I

(211

NHZ (20)

I

ys-----=-=c::..:Cys·Asp·OH

1

5

5

I 5

I 5

I (7) (l9I I Cys'--'-------'----' CyS

(I)

t\~Z'Phe

3

Na::0 Na 5 0

Z 4 6

I

(30I

Ala• 0" rr

INSULIN

)1. -5H 2. OZ. pH 8.5 (21)

SO; (I)

SO;

(6) I

(7)

(II) I

NHZ (20)

I

NHZ'Gly ---Cys'Cys'--'---Cys---Cys'Asp'OH I _ I _ 503 A-CHAIN 5°3'

+

SO; (J)

NHZ' Phe

I

Cys

B-CHAIN (7)

SO; (19)1

Cys

(30)

Ala' OH

Fig. 3.-Resynthesis of insulin from its separated A- and ll-chains.

and associates in Canada and a group of Chinese investigators in Shanghai almost simultaneously reported the reduction of insulin to its two chains, the separation of the chains in' the S-sulfonate form and subsequently regeneration of insulin by oxidation of a mixture of the sulfhydryl forms of the two chains. One of the two groups also reported isolation of crystalline insulin regenerated in this way. This overall process is illustrated in figure 3. Insulin is treated with sodium sulfite and sodium tetrathionate. As a result of this treatment, the disulfide bridges are broken and the liberated sulfhydryl groups arc converted to the S-sulfonale form. The chains, in the S-sulfonate form, are separate, purified and on treatment with a thiol, such as mercaptoethanol or thioglycolic acid, arc converted to the sulfhydryl form. On airoxidation of a mixture of the sulfhydryl forms of the two chains, insulin is regenerated. Thus if the S-sulfonate form of the A and B chain's could be synthesized, insulin synthesis should be achieved by a similar series of reactions. As was stated earlier, the only practical approach for the preparation of polypeptide chains of such length is by "fragment condensation," that is by synthesizing, stepwise, polypeptide subunits and then' attempting their condensation to form large peptide fragments. This approach was implemented by first undertaking the preparation of subunits, slarting from the C-terminal end of the A-chain. I may mention here that all the synthetic peptldes that I am going to discuss have been obtained in chemically pure form as judged by elemental analysis, paper chromatography and in certain cases by qunntitative amino acid analysis. The stereochemical homogeneity of several of these synthetic polypeptides has been established by incubation' with leucine aminopeptidase. Complete hydrolysis of the synthetic products by this enzyme to the constituent amino acids was taken as proof that the optical purity of the containing amino acids was preserved during the synthetic processes. In figure 4 the synthesis of the

1064

}'. G. KATSOYA!\'lIiIS

NH;t I I.HBr/AcOH Z'Asp'ONBz 2. Z'Cys.PNP I Bz

Bz NH;t

I

I

z. Cys-Asp ' ONBz

I. HBr/AcOfI

2. Z'Tyr. PNP I OBz

Bz N~ I I I. HBr/AcOH Z'Tr-CYS-ASP 'ONBz 2. Z'Asp 'PNP I OBz NH

NH;t Bz NH;t I I I Z · Asp-Tyr-Cys-Asp 'ONBz

z

I. HBr/AcOH 2. Z'Glu 'PNP I O-OBz

e-os,

NH;t Bz NH;t I I I I Z- Glu- Asp- Tyr-Cys-Asp •ONBz

Fig. 4.-Synthesis of fragment A17 -2 1• NH2 I

H·Leu·OMe

I. HBr/AcOH ~

2. Z'Tyr 'PNP

z- Glu' PNP

..

NH2 I

z- ere- Leu' OMe

Bz NH2 I I Z· Tyr' Glu·Leu· OMe

I

I. HBr/AcOH

2. z-Leu' PNP

~

Bz

NH2 I Z· Leu.Tyr 'Glu • Leu· OMe

NH2 I Z·Leu·Tyr .Glu·Leu.NHNH 2

Fig. 5.-Synthesis of fragment A13 - 1G• C-terminal pentapeptide fragment is described. Starting with Cbzo-Asp (NII 2 ) • ONBz, we removed the Cbzo-protecting group with HBr/AcOH and the resulting asparagine p-nitrobenzyl ester was condensed with Cbzo-S-BzCys-PNP to give the dipeptide Cbzo-S-Bz-Cys-Asp (NH 2 ) • ONBz. By a similar series of reactions we obtained eventually the C-terminal pentapcptide fragment Cbzo-y-Bz-Glu-Asp (NH 2)-Tyr-S-Bz-Cys-Asp (NH 2 ) • ONBz. Protection of the carboxyl group of asparagine presented a major difficulty in our earlier studies. Conventional methods of protection, namely, by conversion to a methyl or ethylester, could not be applied since alkaline saponification of asparagine esters gives rise to a cyelic imide and eventually to a mixture of asparagine and isoasparagine. Likewise, conversion to a benzyl ester could not be used since such esters are labile to IIBrI AcOH treatment. TIle use of the IJ-nitrobenzyl group circumvented these difficulties, because this group is stable to HBrl AcOH treatment (an obligatory process for de-

1065

PROBLEM OF CHEMICAL SYNTHESIS OF PROTEINS

y-08z NH2 5z NH2 I I I I Z· Glu·Asp·Tyr 'Cys'Asp'ONBz

NHZ I

Z· Leu 'Tyr' Glu-Leu- HNNHz

!

!

HBr/AcOH

NH2 I Z· Leu ·Tyr • Glu- Leu' ON3

y-OBz NH2 Bz NH2 I I I I H' Glu- Asp·Tyr • Cys'Asp' ONBz

I_-l-_I NHZ I

y-OBz NHz I I

z· Leu 'Tyr • Glu' Leu' Glu>Asp ·Tyr

!

Bz NHZ I I • C)'s ·Asp· ONBz

HBr/AcOH

NHZ y-OBz NHZ Bz NHZ I' I I I I H' Leu .Tyr • Glu· Leu· Glu·Asp ·Tyr • Cys·Asp· ONBz

Fig. 6.-Synthesis of fragment Al 3 .:!l ' Bz I

Z'Cys'PNP

+

H'Ser'OMe

~O

STEPS

Bz I

Z'VOI'Pycys.ser'OMe

Bz I

Z'Vol'Cys'Ser'OMe

!

HzN'NH z

~HZ r-{)~z ~Hz

Bz I

Bz

~z ~Hz

z- Leu·Tyr·Glu·Leu·Glu·Asp·Tyr·Cys·Asp·ONBl

Z·Vol'Cys.Ser·HN·NA z

!

!

I

Z·VaI·Cys·Ser·ON3

I

+

~HZ r'O~z ~H2 ~z ~H2 H·Leu·Tyr·Glu·Leu·Glu·Asp·Tyr 'Cys'Asp'ONBz I liN TWO STEPS

~z ~H2 r'O~z ~H2 ~z ~H2 H'Vol' Cys'Ser' Leu' Tyr-Glu' Leu·Glu·Asp-Tyr• Cys'Asp'ONSl

Fig. 7.-S)'nthesis of subunit A10 . 2 1 ' blocking cysteine-containing pep tides ) and can be removed by Na/NH 3 treatment. In figure 5 the synthesis of the tetrapeptide sequence next to the pentapeptide described in figure 4 is shown. Here also the stepwise approach is employed. The Cbzo-pentapeptidc, whose synthesis was described in figure 4 was decarbobenzoxylatcd with HBr/AcOH. The tetrapeptide hydrazide, whose synthesis was described in figure 5, was converted to the azide. Coupling of these two fragments resulted in the formation of the nonapeptide fragment which occupies the C-terminal sequence of the A-chain (Hg. 6). The decarbobenzoxylated C-terminal nonapeptide was condensed with the azide of the tripeptide Val-Cys-Ser which is adjacent to the nonapeptidc in the A-chain. The

1066

1'. G. KATSOYAKKIS

~/Pd-C

y-?But Z · Glu· OMe

y-?But I. Z ' Vol , PNP • • Glu : OMe - - - - - - 2. H2/Pd-C

y-OBut y-OBut . I l Z . lieu . PNP I H,VOI-Glu .OMe2.H2/Pd_C • H·lIeu-Vol-Glu 'OMe

y-OBut

I

p-NO ·Z ·Gly·PNP 2 •

OW

p- N02 . Z . Gly -lieu - Vol- Glu . OMe

y-OBut I P - NOz ' Z . Gly-I1eu -Val-Glu 'OH

Fig. B.-Synthesis of fragment A1-4' 8z I I. Z ·Ala · PNP ~H'Ala-Gly' OEt I. Z - Cys . PNP H . Gly . OEt 2. HBr/ AcOH 2, H8r/AcOH

8z t H 'Cys-Ala-GIY'OEt NHZ I Z 'Glu'PNP

r

z 8z 8z I Z ' CyS . PNP I I H8r/AcOH· H'Cys-Cys-Ala-GIY 'OEt

i

NHz 8z 8z I f I HBr/AcOH Z-Glu-Cys-Cys-Ala-GIY 'OEt •

~H2 ~z

~z -

H'Glu-Cys-Cys-Ala-Gly' OEt

Fig. 9.-Synth esis of fragment A5 -9 •

final product of this condensation was the dodecapeptide which contains the C-terminal sequence of the A-chain (fig. 7). \Ve then started working on the N-terminus of the A-chain. In figure 8 the synthesis of the N-terminal tetrapeptide is described. You may notice that the y-COOII of glutamic acid is protected by conversion to ·a tert-butyl ester. This ester is stable in hydrogenolysis and stable to alkaline treatment. Thus, usin'g catalytic hydrogenation to deblock after each stepwise addition of amino acid residues and alkaline saponification to selectively unmask the a-carboxyl function of glutamic acid, we were able to prepare the partially protected tetrapeptide se
1067

l'TIODLEM OF CIlE:\UCAL SYl'\TIlESIS OF PTIOTEIXS

y-08uf I p- N02"Z" Gly"Ileu" Val" Glu> OH

+

NH2 8z 8z I I I H - Glu> Cys - Cys -Ala" Gly' OEt

~ !

y-08uf NH2 8z 8z I I I I P-N02-Z- Gly-Ileu -vcl - Glu - Glu - CyS" Cys -Ala - Gly -OEI

OH-

y-08uf NH2 8z 8z I I I I p- N0 2"Z -Gly-Ileu - Val - Glu - Glu -Cys -Cys -Ala -Gly - OH

Fig. IO.-Synthesis of subunit

AI.f)'

7,OBul NHz Bz Bz I I I I P,NQz.z.GIY'Ileu.VOI'GIU'GIU'CYS'CYC]2.Alo.GIY'OH Bz NHz 7'OBz NH z Bz NHz I I I I I I H·Vol, Cys -Ser- Leu·Tyr· Glu' Leu·Glu ·Asp·Tyr·Cys ·Asp·ONBz

J d18ul NHZBz Bz Bz NHz 7,OBz NHz 8z NHz I I I I I I I I I I p-N0Z' t. Gly'I1eu ,Val' Glu' Glu' Cys'Cys'Alo' Gly·Val·Cys 'Ser' Leu 'Tyr -Glu -Leu-Glu ·Asp·Tyr-Cys·Asp·ONBz

I. HBr/CF3COOH 2. Na/NH3 3. NO zS03 + NO ZS406

~~~ I I I

~ I

~ I

~ I

~~z I ~ J

H·Gly'Ifeu-Val, Glu' Glu' Cys·Cys·Alo·Gly·Val· Cys·Ser·Leu ·Tyr· Glu' Leu·Glu· Asp·Tyr 'Cys ·Asp·OH

Fig. H.-Final step in the synthesis of the A-chain of insulin.

terminal nonapcptide of the A-chain (fig. 10). \Ve had thus in our hands two subunits, the N-terminal nonapeptide and the C-terminal dodecapeptide, which contain all the amino acid sequences of the A-chain of insulin. The next step was to condense these two subunits in an attempt to obtain the A-chain. This indeed was done as shown in figure 11. The nonapeptide was condensed with the dodecapeptide by the carbonyldiimidazole method to give the fully protected sulfhydryl form of the A-chain of insulin. \Vithout any further purification the protected product was treated successively with a. HBr/CFaCOOH to remove the lcrt-butyl and· p-nitrocarbobenzoxy groups; b. Na/NII 3 to cleave the S-benzyl and p-nitrobenzyl groups; c. Na:!SOa + Na:!S40C to convert the liberated-SH to the S-sulfonates. The deblockcd material was purified by column chromatography. The S-sulfonatc form of the A-chain was contained in a single peak. Amino acid analysis of this synthetic material after acid hydrolysis gave a composition, in molar ratios, which is consistent with the theoretically expected values for the Achain.

1068

P. G. KATSOYANKIS

Tos I Tos I Z 'Lys op NP I. H8r/AcOH Alo ' OMe -----'----Z·Lys-Alo 'OM e 2.Z· Pro ·PNP Tos Tos . I I. Hz!Pd-C I Z· Thr-Pro-Lys-Alo- OMe Z· Pro-Lys-Alo OMe 2. ZoThr + DCC 0

I. ~/Pd-C 2. Z . Tyr . PNP I 8z

8z

Tos

I

I I. Hz/Pd-C Z · Tyr-Thr-Pro-Lys-Alo OMe ~2:"": . Z=-'oo.:.,p~h'::"e-: o P~N-::P0

Tos

I. H /Pd-C Phe . PNP 3. Hz/Pd-C Tos I H · Phe- Phe-Tyr-Thr-Pro-Lys-Alo oOMe J

Z 'Phe-Tyr-Thr-Pro-Lys-Alo ' OMe

2.

z·

Tos: tosyl

Fig. 12.-Synthesis of fragment B2 t .: O' Tos Z'GIY'PNP

+

I H'Phe' Phe·Tyr·Thr· Pro'Lys'Ala' OMe

~TWO .. -Tos

I Z'ArQ'OH

Tos I H' Gly' Phe' Phe ·Tyr·Thr· Pro' Lys 'Alo' OMe

+

~TWO r-08z

STEPS

STEPS

",-Tos'

I

Z· Glu- PNP

+

Tos

I

I

H' Arg' Gly' Phe' Phe' Tyr' Thr' Pro' Lys 'Alo' OMe

~

y-08Z ...-Tos Tos I I I z-ero- Arg' Gly' Phe' Phe 'Tyr ·Thr· Pro· Lys'Alo' OMe

!

H8r/AcOH

y-08z "'Tos Tos I I I H'Glu 'Arll' Gly' Phe' Phe 'Tyr' Thr· Pro' Lys'Alo' OMs

Fig. 13.-Synthesis of fragment B2 1•3 0 •

When this material was subjected to electrophoresis on a cellulose-supporting medium, the product obtained had the same amino acid composition as the starting preparation and gave a sharp single Pauly -positive spot on ascending and descending paper chromatography and a sharp single Pauly-positive spot on high voltage paper electrophoresis. . A final indication that our synthetic product is the A·chain of insulin came from recombination experiments between the synthetic A-chain and the

pnODLE~[

Pl\OTEI~S

OF CHEMICAL SYNTIIESIS OF

Oz

Bz

I

Z'Cys'PNP

H'Gly'OEI

Bz

I.HBr/AcOH 2. Z'Tyr-PNP

I

Z· Cys'GIY'OEI

Bz

I

Z· Leu·Yol,Cys·Gly·OEI

Bz

I

~

I. HBr/AcOH 2. Z'Vol'PNP

Bz

I. HBr/AcOH 2. Z'Leu'PNP"

I

Z· Yol'Cys'GIY'OEI

1069

I

Z· Tyr' Leu 'Vol' Cys' Gly'OEI

I.HBr/AcOH ~ 2. Z·leu·PNP

Bz Bz I

Z»Leu'Tyr ' Leu-vet- Cys'Gly' OEI

I.HBr/AcOH 2 . Z'Alo'PNP

Bz I

Z·Alo· Leu·Tyr· Leu' Yol'Cy s' Gly· OEI

..

ow

Bz I

Z·Alo· Leu'Tyroleu'Yol'Cys 'GIY'OH

1"ig. 14.-Synthcsis of fragment B14.:!o' Bz I Z· Ala'Leu'Tyr' Leu'YoHys' Gly' OH

cor .-------' Tos OR r -OBz e-tes Dee H'Glu' Arg· Gly' Pbe -Phe ·Tyr·Thr-Pro ·Lys· Ala· OMe I

I

I

!

Bz r-OBz w·Tos Tos I I I I Z·Alo'Leu'Tyr' Leu 'Vol'Cys'GlY 'Glu'Arg· Gly'Phe 'Phe ·Tyr ·Thr ·Pro ·Lys'Alo 'OMe

1HBr/AcOH Bz r -OBz w-Tos Tos I I I I H'Alo'Leu -Iyr-Leu·Yol· Cys'Gly' Glu- Arg -Gly-Phe- Phe·Tyr ·Thr·Pro -Lrs ·Ala 'OMe Fig. 15.-::-Synthcsis of subunit B U •3 0 '

natural B-chain. These results also are strong evidence that the structure proposed by Sanger for the A-chain is correct. For the recombination experiments, which were carried out by Dr. Dixon and Dr. \Vilson of the University of Toronto, we used the synthetic material obtained by column chromatography, without further purification, and natural B-chain prepared from ox insulin. Insulin activity was generated in a yield ranging from 0.5-1.2 per cent. The activity was determined by the mousediaphragm method and hy immunologic assays. In the latter case the regenerated insulin activity was neutralized by anti-ox-insulin serum. \Ve have thus accomplished our first objective, namely the synthesis of the A-chain of insulin. \Vhile the work on the A-chain was in progress, synthetic studies on the B-chain were well under way. This story is given in the next few figures. In

1070

P. C. KATSOYANNIS

figure 12, the synthesis of the C-terminal heptapeptide, corresponding to positions 24 to 30 in the B-chain, is illustrated. In all but one step, the p-nitrophenyl esther method was applied for chain elongation. Alanine methyl ester was condensed with Nv-Cbzo-Ne-tosyl-lysine P: nitrophenyl ester and the resulting crystalline dipeptide was decarbobenzoxylated on' exposure to HBr in acetic acid and coupled with the P: nitrophenyl ester of Cbzo-proline to give the protected tripeptide. Catalytic hydrogenation of Cbzo-Pro-Lys-Ala.OMe over palladium on charcoal and coupling of the ensuing deblocked product with Cbzo-threonine by the dicyclohexylcarbodiimide method afforded the tetrapeptide Cbzo-Thr-Pro-Lys-Ala.OMe. By a comparable series of reactions, the tetrapeptide was converted to the pentapcptide, then hexapeptide and finally to heptapeptide. In figure 13 the synthesis of the C-terminal decapeptide is described. Its preparation' was accomplished by stepwise elongation of the amino-free heptapeptide ester whose synthesis was described in figure 12. The glycine residue was introduced by the J)-nitrophenyl ester method, the arginine residue, with the guanido function' protected with the p-toluenesulfonyl group, was introduced by activation with a new peptide reagent, the 2-ethyl-5-pheuyloxazolium-3'-sulfonate, and the glutamic acid residue was attached by the p-nitrophenyl ester method. In figure 14 the synthesis of the heptnpeptide sequence adjacent to the decapeptide just discussed is illustrated. It is a straightforward stepwise elongation of a peptide by the J)-nitrophenyl ester method. This heptapeptide, which occupies positions 14-20 in the B-chain, was then condensed with the decapeptide which occupies positions 21-30 of that chain as shown in figure 15. Product of this condensation is the heptadecapeptide which contains the C-tenninal 17 amino acids of the B-chain (sequence 14-30). It is remarkable that up to now all of our synthetic polypeptides were purified by conventional technics, namely precipitation or sometimes crystallization from organic solvents and water. This offers the great advantage that we are in position to prepare such compounds in gram-quantities, a fact that it would not be possible if physical methods of purification, such as countercurrent distribution and chromatographic procedures, were necessary. Next we turned our efforts to the synthesis of the remaining sequence of the B-chain, namely the N-terminal tridecapeptide (sequence 1-13). The synthesis of this peptide is shown in the next few figures. In figure 16 the synthesis of the partially protected N-terminal pentapcptide sequence (sequence 1-5) is illustrated. Starting with im-benzyl-histidine benzyl esterand using the p-nitrophenyl ester method for activation, this peptide was also prepared by the step' wise elongation approach. The synthesis of the tetrapeptide adjacent to the pentapeptide just described and occupying positions 6-9 is shown in figure 17. In this case also, the stepwise approach was applied using the J)-nitrophenyl ester method for activation. Coupling of the two fragments, namely the pentapeptide (sequence 1-5), and tetrapeptide (sequence 6-9), afforded the N-terminal nonnpeptide of the

pnOBLE~[

OF

CHE~nCAL SYNTHESIS

10il

OF PHOTEI1'\S

NHz I Z'G lu -PNP

+

8z I His ' 08z

~ostep3

NHz I Z'Asp ·PNP

+

NHz Bz I I H-Glu-His 'OBz

~~OSlepS_ NHZ NHz Bz I I I H-Asp-Glu-His ' OBz

+

Z' vet - PNP

~steps_

Z· Phe . PNP

NHz NHz Bz I I I H'Val-Asp-Glu-His' OBz

+

~
NHz NHz Bz I

I

I

Z· Phe-Vol-Asp -Glu -His- Of I

Fig. 16.-Synthesis of Fragment Rt .:;.

+

Z·Gly·PNP

H'Ser',OMe

~

TWO STEPS

8z I

Z·Cys·PNP

+

H'GIY'Ser'OMe

~

TWO STEPS

8z

+

Z -Leu'PNP

I

H-Cys'Gly-Ser'OMe

T

8z I Z- Leu' Cys' Gly' Ser-OMe

~

H8r/CF3COOH

8z I II' Leu- Cys' Gly·Ser- OMe

Fig. 17.-Synthcsis of fragment Bc. !) .

B-chain (sequence 1-9) as shown in figure 18. The nonap cptide ester was then converted to the corrcsponding hydrazide. In figure 19 the synthesis of the remaining tetrapeptide, namely scqucncc 10-13, is illustrated. In this case, because of the nature of the constituent amino acid residues, we followed a somewhat different approach. We prepared two dipeptide derivatives, CbzoBz-His-Leu-OMe and Val-Glu-r-Onut and by their condensation thc fully protcctcd tetrapeptide was obtained. Selective saponification of the a ester function gavc us the partially protected tetrapeptide.

1072

P. G. KATSOYANNIS

NH2 NH28z I I I Z· Phe ·Yal·Asp· Glu- His- OH

8z I H· Leu' Cys' Gly' Ser- OMe

+

~

z·

NH2 NH2 Bz Bz I I I I Phe ·Yal • Asp· Glu- His' Leu-Cys- Gly' Ser • OMe

!

H2N'NH 2

NH2 NH2 Bz I I I

Bz I

Z· Phe -vcr-Asp- Glu' His' Leu-Cys- Gly' Ser > NHNH 2

Fig. lB.-Synthesis of fragment BI _!) .

Z'Yal'PNP

+

,.-OBuI I H'Glu'OEt

T

r- OBu1

I Z'Yal'Glu'OEt

~ HZ/Pd Bz I Z·His·Leu·OH

,.-OBuI I H·Val·Glu·OEI

+

I

I

JDee Bz

,.-OBul

I

I

Z· His·Leu·Yal·Glu·OEI

~

OH-

Bz ,.-OBuI I I Z'His' Leu·Yal·Glu·OH

Fig. 19.-Synthesis of fragment BI O-13 '

The final step in the synthesis of the N-terminal tridecapeptide, sequence 1-13 of the B-chain is shown in figure 20. Deblocking of the tetrapeptide, whose synthesis was illustrated in figure 19, was carried out by controlled catalytic hydrogenation. The ensuing deblocked tetrapeptide was condensed with the azide of the N-terrninal nonapeptide. 11w azide of this peptide was prepared from the corresponding hydrazide, whose synthesis was illustrated in figure 18. The synthesis of this tridecapeptide completes the other major subunit needed for the preparation of the B-chain. What is left to complete the synthesis of the B-chain is the coupling of the partially protected C-terrninal heptadecapcptide and N-terrninal tridecapeptide and the removal of all the protecting groups from the reaction product. Preliminary experiments to that effect are now underway. As shown in figure 21 we condensed the tridecapcptide and heptadecapeptide subunits by the

PHOnLE~I

1073

OF CHEMICAL SYKTIlESIS OF PHOTEINS

~H2 ~H2 ~Z

~Z

Bz

r'08ut

I

!

!

H2/Pd

HN02

~H2 ~H2 ~z

I

Z· His·leu·Val·Glu·OH

Z· P/'Ie·Val·Asp· Glu 'His 'leu 'Cys 'GIY'Ser- HNNH 2

~l

Bz

Z· Phe 'Val'Asp' Glu- His 'leu' Cys' Gly 'Ser'ON~

+

r-OBu

I

I

H' His ·leu·Val· Glu- OH

I +

NH2 NH2 Bz Bz 8z r-OBut I I I I I I Z· Phe 'Val'Asp 'Glu' His 'leu' Cys 'GIY'Ser' His-leu 'Val'Glu' OH

Fig. 20.-Synthesis of subunit Bl - 13 N~ I

NHzBz I I

Bz I

Bz I

T-OBul ,

z· PIle -Yal- Asp' Glu' His'Leu' Cys'Gly' Ser- His'LeuVol' Glu 'OH 8z T-ORz "'-Tos Tos I T I I H·Ala' Leu·Tyr· Leu-Yol-Cys- Gly'Glu-Arg' Gly-Phe-Phe'Tyr -Thr'Pro-Lys -Ala-OMe

N~ NHZBz

~

8z Bz y-OBuI 8z y-08z "'.Tos I I I I I I Z· Ptie ·Yal-Asp' Glu'His' Leu- Cys' Gly -Ser- His- Leu-Yo I·Glu -Ala -Leu 'Tyr' Leu'Yal'Cys' Gly -Glu'A!g Gly "rt

I

I

los I

!'he •

MeO'Ala- Lys'Pro·Thr -Tyr' Phe I. Cf3COOH

2. Na/NH 3 NH2 NH2 I I

SOi I

3. No zS03 + N02 $4°6, OW SOi T

H· PIle· Yal· Asp' Glu' His'Leu' Cys' Gly· Ser· His' Leu·~I· Glu'Ala· Leu ·Tyr· Leu 1101· Cys' Gly 'Glu 'A!g G!y

PIle HO·Alo·Lys·Pro·Thr·Tyr·Phe

Fig. 21.-FinaI step in the synthesis of the Bvchain of insulin.

N,N'-carbonyldiimidazole method. The resulting product was deblocked on e>;posure to trifloroacetic acid and to soduim in liquid ammonia and then treated with sodium sulfite and sodium tetrathionate to convert the sulfhydryl groups to the S-sulfonate form. The deblocked material was submitted to column chromatography. A product was thus obtained which, without any further purification, was subjected to combination experiments with the A-chain. As judged by biological assay, using the mouse-diaphragm method, insulin activity was generated when this product was combined with A-chain. However, the yield of insulin activity generated was very small in' comparison with the activity obtained when natural B-chain is recombined with A-chain. \Ve

1074

1'. G. KATSOYANNlS

realize that the synthetic B-chain preparation is not yet pure. It should be pointed out, however, that under these conditions of an impure preparation, the yield of activity is not necessarily a direct measure of the amount of B-ehain present. In fact, thc amount of regenerated activity may be taken as a minimal figure, since the impurities present may modify the condition's of recombination so that optimal recombination of the chains is not obtained. In any case, 'we consider it impractical at the present time to attempt to isolate synthetic insulin by recombination of the crude preparation of the B-chain and the A-chain because of the low yield obtained. Further studies on' the purification of the synthetic B-chain are now underway in our laboratory and we hope that a more purified product may eventually be obtained and thus make possible the isolation of chemically synthesized insulin. ACKNOWLEDGMENT In this work the author had the collaboration at various time intervals of Drs. A. Tometsko, K. Fu-Kuda, K. Suzuki and 11. Tilak. Present address: Division of Biochemistry, Medical Research Center, Brookhaven National Laboratory, Upton, L. I., New York.